This article is also available in Russian.
The Office of National Statistics (ONS) of the United Kingdom has released a provisional analysis of deaths related to COVID-19, as reported on the death certificate, from different ethnic groups in England and Wales. The analysis was prepared using linked census and mortality records on deaths occurring between 2 March and 10 April (reported as of 17 April 2020). While ethnicity is not recorded on death certificates in England and Wales, the researchers linked the deaths involving COVID-19 to the 2011 Census. The 2011 Census data included self-reported ethnicity.
Results of the preliminary analysis show that the risk of death involving COVID-19 among some ethic groups in England and Wales is significantly higher than that of those of white ethnicity. The analysis also shows that socio-economic disadvantage and other circumstances can partly account for some of the difference in COVID-19 mortality between ethnic groups, although the remaining difference has not yet been explained.
The researchers used binary logistic regression models to estimate whether the risk of dying from COVID-19 is greater among ethnic groups than among the white ethnic population. The models controlled for geographic, demographic, socio-economic, living arrangements, and health as measured in the 2011 Census. The report explains the justification for each of the variable used as controls in the model. Detailed information on the methodology used in the analysis is also available online.
After adjusting for socio-economic and demographic factors, the main findings show that most minority groups tend to be more disadvantaged than their white counterparts, and that a substantial part of the difference in COVID-19 mortality between ethnic groups is explained by socio-economic circumstances.
Some of the strengths of the analysis include a high rate of linking of death and census information, and the relative robustness of the data. On the other hand, there is no direct measure of emigrations, which potentially introduces bias and could lead to underestimating mortality. Two other important limitations are the possible delay in registering deaths and dated information on the socio-demographic factors based on the 2011 Census. Thus, ONS is planning to repeat the analysis and to prepare more detailed age-standardized mortality rates of death as more data become available.